Research A 200 Year Record of Atmospheric Cobalt, Chromium, Molybdenum, and Antimony in High Altitude Alpine Firn and Ice KATJA VAN DE VELDE,† C H R I S T O P H E F E R R A R I , †,‡ C A R L O B A R B A N T E , §,| I V O M O R E T , §,| TANIA BELLOMI,§ SUNGMIN HONG,† AND C L A U D E B O U T R O N * ,†,⊥ Laboratoire de Glaciologie et Ge´ophysique de l’Environnement du CNRS, 54 rue Molie`re, B.P. 96, 38402 Saint Martin d’He`res, France, Institut des Sciences et Techniques de Grenoble, Universite´ Joseph Fourier de Grenoble, 28 Avenue Benoıˆt Frachon, B.P. 53, 38041 Grenoble, France, Dipartimento di Scienze Ambientali, Universita di Venezia, Dorsoduro 2137, 30123 Venezia, Italia, Centro di Studio sulla Chimica e le Tecnologie per l’Ambiente-CNR, Dorsoduro 2137, 30123 Venezia, Italia, and Unite´s de Formation et de Recherche de Me´canique et de Physique, Universite´ Joseph Fourier de Grenoble (Institut Universitaire de France), B.P. 68, 38041 Grenoble, France
High altitude cold snow and ice cores from midlatitude mountain ranges have been used very little to obtain historical records of environmental contamination by heavy metals. Co, Cr, Mo, and Sb have been measured by DF-ICP-MSMCN (double focusing inductively coupled plasma mass spectrometry with microconcentric nebulizer) in various sections of a 140 m snow/ice core drilled at a high altitude location near the summit of Mont Blanc in the FrenchItalian Alps. The bottom of the core is older than 200 years. It gives the first snow and ice time series for these metals of high environmental interest for the post Industrial Revolution period. Measured concentrations range from 26 to 433 pg/g for Co, from 8 to 469 pg/g for Cr, from 0.2 to 50 pg/g for Mo, and from 0.2 to 109 pg/g for Sb. For all four metals, concentrations in recent snow are found to be, on the average, significantly higher than concentrations in ice dated from before the middle of the 19th century. There are however differences in the timing and the amplitude of the observed increases from one metal to another. Mo shows the greatest increase (×16), followed by Sb (×5), and Co and Cr (×2-3). Contribution from natural sources is, on the average, limited except for Mo in ice dated from before the middle of the 19th century. For recent snow, contribution from oil and coal combustion is the dominating source for Co, Mo, and Sb. For Cr, on the other hand, the most important contribution is from iron, steel, and ferroalloy industries.
1. Introduction The study of the frozen archives which can be found in the Greenland and Antarctic ice caps has given a wealth of fascinating information on the history of the composition of 10.1021/es990066y CCC: $18.00 Published on Web 09/14/1999
1999 American Chemical Society
the atmosphere of our planet (see, e.g. refs 1-5). Among the most interesting species to be studied are heavy metals such as Pb, Hg, and Cd. Investigation of the changes in the occurrence of these metals in dated snow and ice from central Greenland has for instance allowed for documenting early large scale pollution of the atmosphere of the northern hemisphere for Pb and Cu two millennia ago (6, 7) and the recent history of heavy metal pollution from the Industrial Revolution to present (see, e.g. refs 8-10). Interesting information was also obtained for the southern hemisphere from Antarctica (see, e.g. refs 11-14). On the other hand, very little effort was devoted toward deciphering the history of atmospheric heavy metal pollution on a regional scale in Europe or other temperate areas from the cold snow and ice fields which can be found in high altitude mountain ranges. To our knowledge, the only published data from such areas are for Pb, Cd, Cu, and Zn at a single location in the French-Italian Alps (15, 16). Until now most studies were devoted to Pb, Cd, Cu, and Zn and to a lesser extend Hg. Surprisingly, there are almost no data for other heavy metals such as Sb, As, Se, Cr, Mo, Ag, Au, Co, and V. Moreover, most of the very few published data for these last metals (see, e.g. refs 17-19) are considered to be unreliable because of contamination problems during field sampling and/or laboratory analysis (20). We present here comprehensive data on the changing occurrence of Co (21), Cr (22), Mo (23), and Sb (24) in alpine snow and ice dated from the past two centuries. These metals are of high environmental interest since their atmospheric cycles are now significantly influenced by human activities (25), especially fossil fuel combustion, iron, steel, and ferroalloy production and refuse incineration. They were obtained from the analysis of various sections of a snow/ice core drilled at a high altitude location in the Mont Blanc range at the French-Italian border. Analyses were performed by DF-ICPMS-MCN (double focusing inductively coupled plasma mass spectrometry with microconcentric nebulizer). The data are discussed in the light of available information on the different anthropogenic and natural sources and their variation since the Industrial Revolution.
2. Experimental Section 2.1. Core Drilling and Dating. The samples were obtained as a 10 cm in diameter snow/ice core drilled at an altitude of 4250 m on the east slope of Doˆme du Gou ˆ ter at about 1.5 km northwest of the summit of Mont Blanc (26). The sampling site (45°50′N; 6°51′E) is characterized by a mean annual temperature of -11 °C and a mean annual snow accumulation rate of ∼3.5 m H2O‚yr-1. The bedrock is ∼140 m below the surface. The close off depth is ∼60 m. The core was drilled from the surface down to the bedrock using a stainless steel electromechanical drill coated with Teflon (PTFE). No wall retaining fluid was used. Great precautions were taken to minimize contamination brought to the outside of the core * Corresponding author phone: +33 476 82 4237; fax +33 476 82 4201; e-mail:
[email protected]. † Laboratoire de Glaciologie et Ge ´ ophysique de l’Environnement du CNRS. ‡ Universite ´ Joseph Fourier de Grenoble. § Universita di Venezia. | Centro di Studio sulla Chimica e le Tecnologie per l’AmbienteCNR. ⊥ Universite ´ Joseph Fourier de Grenoble (Institut Universitaire de France). VOL. 33, NO. 20, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3495
during handling and packing: the sections were handled by operators wearing clean room garb and shoulder lengthpolyethylene gloves. They were packed in double sealed polyethylene bags and transported frozen to the laboratory. The upper part of the core (down to ∼110 m) was dated (26) by combining known reference levels from atmospheric nuclear weapon tests and the Chernobyl accident and a glaciological ice flow model. The depths of the reference levels in the core were determined from a continuous depth profile of gross beta radioactivity. Fallout from the Chernobyl accident was found at a depth of 39 m. The 1963 peak from atmospheric nuclear tests was observed at a depth of 90 m. The dating so obtained shows that the ice at 110 m corresponds to the year 1940. The depth-age scale is however not linear, with a strong thinning of the annual layers from the surface toward the bedrock: as an illustration the thickness of an annual layer is ∼7 m at the surface and ∼1 m at 100 m. The precision of the ages so obtained is very good. Preliminary methane measurements show that the ice close to the bedrock is at least 200 years old (J. Chappelaz and R. Blanchard, personal communication). A tentative depth-age scale was then constructed by interpolation, for depths greater than 110 m, assuming that the ice at 136 m was 200 years old. 2.2. Decontamination of the Core Sections. The full cross section was available for heavy metal analysis. Twenty-four core sections (length ranging from 0.5 to 1 m) were analyzed. The depth of these sections ranges from 21.9 to 136.6 m. Despite the precautions taken in the field significant contamination was present on the outside of the core. It was then mandatory to decontaminate each section by chiseling veneer layers in progression from the outside to the inside of the section (27). It allowed for obtaining the central cylindrical inner core of each core section (diameter 5 cm). After the chiseling was completed, the inner core was cut into several successive parts (three parts in most cases, except for one 1 m long section which was divided into five parts). Each veneer layer and central part was melted separately at room temperature in acid cleaned low-density polyethylene (LDPE) bottles inside a clean laboratory (28). They were acidified with ultrapure nitric acid from the U.S. National Institute for Standards and Technology (29) to make 0.5% solutions. Ten milliliter aliquots were poured into acid cleaned LDPE bottles. The bottles were sealed in acid washed inner and outer polyethylene bags. They were transported frozen to the University of Venice for analysis. To establish whether the central part of the core sections was free from outside contamination or not, changes in Co, Cr, Mo, and Sb concentrations from the outside to the inside of each core section were investigated. Examples of such outside-inside concentration profiles are shown in Figure 1. Good plateaus were observed in the central part of the core in all cases which indicates that no external contamination had reached the inner cores. 2.3. DF-ICP-MS-MCN Analysis. Cr, Co, Mo, and Sb were directly measured, by double focusing-ICP-MS using a Finnigan MAT (Bremen, Germany) “Element” sector field mass spectrometer, at the Department of Environmental Sciences of the University of Venice (30, 31). Acidified multielement synthetic standard solutions from Merck (Darmstadt, Germany) were used after dilution at concentrations similar to those expected in snow and ice samples. The high ions transmission, together with a low background signal given by the sector field mass spectrometer, offer sensitivities at and below the pg/g level, often required in the trace element determination in snow and ice. Furthermore the possibility to work at higher resolution than a quadrupole instrument allows most of the analytes to be separated from interferences forming in the plasma. To minimize the amount of sample consumed for the analysis, 3496
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 20, 1999
FIGURE 1. Measured Cr, Co, Mo, and Sb concentrations as a function of radius in one alpine snow section (10 cm in diameter).
TABLE 1: Measured Concentrations by DF ICP MS for Certified Reference River Watera measured concentrations pg/g
Co Cr Mo Sb
DF-ICP-MS-MN
certified value
26 ( 360 ( 30 220 ( 20 140 ( 20
27 ( 3c 300 ( 40 190 ( 10 120 ( 10
2b
a SLRS 3, Riverine reference material for trace metals, National Research Council of Canada, INMS, Ottawa, Canada. b SD (n ) 5). c 95% confidence limits.
a MCN-100 (Cetac Technologies, Omaha, NB) microconcentric nebulizer was used throughout the work. Possible interferences affecting the accuracy in the determination of Co, Cr, Mo, and Sb in Alpine snow and ice were carefully considered; for this reason the following isotopes and relative resolutions as m/∆m (in brackets) were used: 52Cr (3000), 59Co (3000), 98Mo (300), and 121Sb (300). The chosen isotopes were the most abundant and least interference-prone possible. External calibration curves were used for quantification. Repeatability and precision of the measurements were determined on a snow sample, for which the available sample volume was larger than usual, dated 1980 (depth 54.3 m). Metals concentrations in that sample were (in pg/g) 531, 304, 3.1, and 0.45 for Co, Cr, Mo, and Sb, respectively. The relative standard deviations are 3.3%, 4.6%, 6.5%, and 18.6% for Co, Cr, Mo, and Sb, respectively. We also measured certified reference river water (SLRS 3, Riverine Reference Material for trace metals, National Research Council of Canada, INMS, Ottawa, Canada), Table 1. The values agree quite well. In addition, we also measured Al and Na concentrations to provide indexes for the interpretation of the data. They were analyzed with a precision of ∼10% by graphite furnace atomic absorption spectrometry (GFAAS) using a PerkinElmer Aanalyst 100 spectrometer and HGA 800 graphite furnace.
TABLE 2: Comparison between Heavy Metal Concentrations in Snow from Greenland (30) and the Alps (This Work) Dated from 1989 and 1990 for Mont Blanc and from the Early 1990s for Greenland measd concns (pg/g)
Mont Blanca min-max mean Greenlandb min-max mean a
This work.
b
Co
Mo
Sb
127-433 248
9.5-50 19
1.6-49 25
0.65-15.5 5.8
0.08-7 1.6
0.2-4.3 0.9
Summit area in central Greenland (72°20′N, 38°45′W).
3. Results and Discussion 3.1. Character of the Data. Measured concentrations for Co, Cr, Mo, and Sb in 74 depth intervals range from 26 to 433 pg/g for Co, from 8 to 469 pg/g for Cr, from 0.2 to 50 pg/g for Mo, and from 0.2 to 109 pg/g for Sb. To our knowledge, they are the first data ever published for these metals for nonpolar snow and ice. The only reliable data for snow and ice to which our data can be compared are those recently obtained for Co, Mo, and Sb for snow dated from the 1990s collected from a shallow snow pit in central Greenland (30). Table 2 compares these Greenland data with the data we obtained for snow dated from 1989 and 1990. The mean concentrations observed in recent snow at Mont Blanc are much higher than the corresponding values for Greenland (×43 for Co, ×12 for Mo, and ×29 for Sb). It can be seen that the Greenland values are generally 1 or even 2 orders of magnitude lower than the Mont Blanc values. It exemplifies the very different geographical location of Mont Blanc and central Greenland. Mont Blanc is indeed located in Western Europe close to various sources of anthropogenic heavy metals, while central Greenland is located in the remote Arctic far away from sources.
FIGURE 2. Seasonal variations in the concentrations of selected metals and δD‰ in high altitude ice dated from 1960 to 1961 collected near the summit of Mont Blanc in the French-Italian Alps. Depth scale is shown at the bottom, and the isotopic seasons are shown at the top.
3.2. Short-Term Variability. The investigation of the shortterm (seasonal) variability in the concentrations is important to resolve the relative importance of meteorological and source parameters. Previous investigations in snow dated from the 1960s for Al, Bi, Cd, Cu, Mn, Pb, and Zn have shown that the heavy metal content of Mont Blanc snow is determined by both parameters (15). For these metals, pronounced seasonal variations by up to 2 orders of magnitude were observed with the highest concentrations in summer and the lowest concentrations in winter. These variations are largely linked with changes in the vertical structure of the regional troposphere. Winter is characterized by frequent low altitude thermal inversions limiting the transportation of ground emitted pollutants to high altitude mountains areas. The pollutant content of high altitude winter snow is then largely contributed from distant sources. During summer, on the other hand, there are vigorous vertical changes in the troposphere which allow local and regional ground emitted pollutants to reach the high altitude areas. We have investigated short time variations in concentrations of Co, Cr, Mo, and Sb for 2 years of the 1960s: summer 1960 to summer 1961; and winter 1966/1967 to winter 1967/ 1968. The 1960s decade was chosen since it corresponds to elevated heavy metal concentrations (see section 3.3) linked with important anthropogenic inputs. Moreover the annual layers for this period of time are thick enough to allow for resolving the successive periods of the years (1.8 m from summer 1960 to summer 1961 and 2.2 m from winter 1966/ 1967 to winter 1967/1968).
FIGURE 3. Seasonal variations in the concentrations of selected metals and δD‰ in high altitude ice dated from 1967 to 1968 collected near the summit of Mont Blanc in the French-Italian Alps. Depth scale is shown at the bottom, and the isotopic seasons are shown at the top. As illustrated in Figures 2 and 3, pronounced seasonal variations are observed for Co, Cr, Mo, and Sb with high values for the summer months and much lower values for the winter months. In 1960/1961, the summer-to-winter ratio is ∼6 for Co, ∼5 for Cr, ∼34 for Mo, and ∼60 for Sb. For 1967/1968 it is ∼3 for Co, ∼8 for Cr, ∼65 for Mo, and ∼90 for Sb. As will be discussed later (see section 3.5) such very large changes cannot be explained by short-term changes in the VOL. 33, NO. 20, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3497
FIGURE 4. Changes in Co, Cr, Mo, and Sb concentrations in snow and ice deposited at a high altitude site in the French-Italian Alps since the end of the 18th century. The individual data obtained for 74 different depth intervals have been averaged over periods of 1-3 years, giving the rectangles shown in the figure. It allows to compensate for short-term (seasonal) variations and better reveal the long-term time trends. strength of anthropogenic or natural sources or in mesoscale atmospheric transport. As for the other metals these changes are obviously mainly driven by the meteorological parameters mentioned above. Interestingly, the ratios are much larger for Mo and especially Sb than for Co and Cr. These are the two metals for which the crustal enrichment factors are the highest and the anthropogenic contribution the most important. It was however not possible to investigate these shortterm variations for older annual layers. This is because thickness of the annual layers diminishes with increasing depth. It makes it very difficult to resolve the seasons in ice dated from before the 1940s. It is however likely that pronounced seasonal variations do exist for ice dated before the 1940s. The amplitude of the variations might however be
different since the contribution from anthropogenic sources was less important (see section 3.3.). 3.3. Long-Term Changes. Figure 4 shows changes in Co, Cr, Mo, and Sb in Mont Blanc snow and ice during the past two centuries. To isolate the long-term time trends during the two last centuries, we have summed the individual data points over periods of 1-3 years to compensate for the pronounced seasonal variations (see section 3.2). For Co, Cr, and Mo, Figure 4 suggests that the concentrations remained fairly stable until the 1940s but that they then increased significantly during the recent decades (∼×3 for Co, ∼×4.5 for Cr, and ∼×15 for Mo). Finally, the situation appears to be rather different for Sb. Although Figure 4 suggests that the concentrations of this metal are higher in present century snow and ice than in older ice, our data do not allow for clearly identifying the long-term trends for this metal. It would have been interesting to compare these time series with time series obtained in Western Europe from other archives such as ombrotrophic peat bogs or lake sediments. Unfortunately, we were unable to find such data in the literature. The only exception is time series published by Shotyk and co-workers for Sb (32). It was obtained from the analysis of a dated raised ombrotrophic peat bog (e´tang de la Gruye`re: EGR) in the Jura Mountains in northwestern Switzerland. This is ∼150 km north of Mont Blanc. The EGR concentration profile shows low Sb concentration values until the mid 19th century. The concentrations then clearly increased till the 1920s (∼×17). After that, there is a general decreasing trend from the 1920s to present (with however rather high values for the 1950s). Present day values are ∼×5 lower than the 1920s values. A direct comparison between the Mont Blanc and EGR concentration profiles is difficult especially regarding the very large difference in altitude between the two locations (4250 m for Mont Blanc versus 1005 m for EGR). It is however interesting to notice that the Mont Blanc profile also shows some high values in the ∼1920s followed by lower values afterward. 3.4. Contribution from Natural Sources. Table 3 shows the best available estimates of worldwide natural emissions of Co, Cr, Mo, and Sb to the atmosphere (33, 34). Although there are uncertainties in these estimates, they suggest that rock and soil dust is an important source of all four metals. This is especially the case for Co and Cr, for which rock and soil dust represent about ∼2/3 of natural emissions. Another important source is volcanoes, especially for Cr and Sb. We are not aware of any inventories of natural emissions of Co, Cr, Mo, and Sb for Europe. We have estimated the contribution from rock and soil dust from Al concentrations measured in the samples (mean value: 80.5 ng/g; range: 0.6-1180 ng/g), combined with the mean Co/Al, Cr/Al, Mo/Al, and Sb/Al concentration ratio for the upper crust given in ref 35, Table 4. On the average, a large fraction of Cr and especially Mo in Mont Blanc ice dated before ∼1850 was found to originate from rock and soil dust.
TABLE 3: Worldwide Natural Emissions of Heavy Metals to the Atmosphere for Present Day Interglacial Conditions: Relative Importance of the Different Sources, Range and Median Values for Total Emissions from Refs 33 and 34 Co
a
Mo
Sb
67% 16% 5% 2.5% 9.5%
61.5% 34% 1% 1% 2.5%
43% 13% 19% 10% 15%
31% 28% 8% 24% 9%
total emissions (t/yr) range median
690-11000 6100
4500-83000 44000
140-5800 3000
70-4700 2400
Sea salt and biogenic.
3498
Cr
rock and soil dust volcanoes wild forest fires oceansa continental biogenic
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 20, 1999
TABLE 4: Mont Blanc Snow and Ice Dated from the Last Two Centuries: Co, Cr, Mo, and Sb Contributed from Rock and Soil Dusta Co premiddle 19th centuryb mean range recent decadesc mean range a
Cr
Mo
Sb
8.9 1.5-16
32 6-65
67 21.5-100
3.4 0.8-11.5
8.7 0.1-53.5
21 0.2-91
17 0.4-100
3.7 0.03-82
Expressed in percent of measured concentrations.
b
Samples dated from before ∼1850. c Samples dated from 1960 onward.
FIGURE 5. Crustal enrichment factors (the ranges are shown by bars) for recent snow (from 1960 till 1990) and ancient ice (deposited before the middle of the 19th century) from Mont Blanc. Crustal enrichment factors (EFcrust) have been calculated for each sample and each metal X as follows:
EFcrust(X) )
(X/Al)snow or ice sample (X/Al)mean crust FIGURE 6. The circles represent ice deposited before 1940 and the triangles snow and ice deposited after 1940, with regression lines for each set of data: (top) Cr/Sb concentration ratios and (bottom) Mo/Sb ratios.
where (X/Al)snow or ice sample and (X/Al)mean crust are the concentration ratios of element X to Al, respectively, in the snow or ice sample and in the mean crust, see ref 35. Figure 5 shows the ranges and mean values for EFcrust for two different periods of time: (a) the period before the middle of the 19th century and (b) the period after 1960 for the four metals. With the exception of Mo in premiddle 19th century ice, mean EFcrust values range from ∼6 to 250. It confirms the fact that a large fraction of these metals in Mont Blanc snow and ice originate from other natural sources and/or anthropogenic sources. After correction for the contribution from rock and soil dust, Na concentrations measured in the samples (mean: 4.3 ng/g, range: 0-47.5 ng/g) provide an index for estimating Co, Cr, Mo, and Sb contributed from sea salt spray. We have used the mean X/Al ratios in present day bulk sea water (2.9, 1.8, 0.9, and 33 × 10-9 for Co, Cr, Mo, and Sb, respectively (36, 37)). Contributions so calculated are extremely low (ranging from 0 to 1.4 fg/g). Even if combined with possible natural enrichments of these metals in sea derived aerosols relative to bulk seawater (33, 36, 38) they remain negligible.
Our data do not allow for evaluating the contribution from other possible natural sources such as volcanoes, wild forest fires, and continental biogenic emissions. It is however likely that these last contributions cannot entirely account for the large excesses above rock and soil contribution observed for Co, Cr, and Sb and to a lesser extent for Mo. The observed excesses are largely due to anthropogenic inputs with a possible exception for Mo in the older samples. 3.5. Contribution from Anthropogenic Sources. Table 5 shows estimates of Co, Cr, Mo, and Sb anthropogenic emissions in Western Europe for 1979 (39). They are the only data we could find from available literature: we are indeed not aware of emission inventories for these metals for Europe for other periods of time. Also, we were unable to find any published information about the history of these emissions during the last two centuries.
TABLE 5: Anthropogenic Emissions of Co, Cr, Mo, and Sb to the Atmosphere in Western Europea (1979 Values): Relative Importance of the Different Source Categories and Total Emissions from Ref 39
a
Co
Cr
Mo
Sb
39% 60%
6.5% 8.5% 3.5% 81.5%
42% 58%
32% 40.5%
19000
850
oil and coal fired power plants fuel combustion cement production iron, steel, and ferro-alloy manufacturing refuse incineration
1%
total emissions (t/yr)
2000
26.5% 400
Western Europe: Belgium, France, Germany, Italy, Luxembourg, Portugal, Spain, Switzerland, and The Netherlands.
VOL. 33, NO. 20, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3499
TABLE 6: Ratios for Metal Emissions in 1979 in Western Europea (39) and Mont Blanc Snowb Dated from the Late 1970s, This Work emissions snow
Cr/Sb
Co/Sb
Mo/Sb
Co/Cr
Mo/Cr
Co/Mo
36 0.2
3 0.5
1.1 0.2
0.09 2.4
0.03 0.8
3 3
a Western Europe: Belgium, France, Germany, Italy, Luxembourg, Portugal, Spain, Switzerland, and The Netherlands. b Metal concentrations in excess above rock and soil dust contribution.
Table 5 shows that oil and coal combustion is by far the dominating source for Co, Mo, and Sb (∼100% for Co and Mo and ∼70% for Sb). For Cr, on the other hand, the dominating source is iron, steel, and ferro-alloy manufacturing (∼80%). A key parameter for the atmospheric transport of metals from source areas is the size distribution of the aerosols in which these metals are found. Interesting information on that distribution is given by Quinn and Ondov (40). These authors have studied the size distribution of various metals, including Co, Cr, Mo, and Sb at ground level in the Chesapeake Bay area in Eastern United States. They found that Sb was present in very small size particles with a mass median aerodynamic diameter (MMAD) of ∼0.3 µm. On the opposite, a significant fraction of Cr was found on much coarser particles with a MMAD of ∼10 µm. The situation was intermediate for Co and Mo, with MMAD values of ∼5 and 6 µm. It suggests that Cr bearing particles are likely to be deposited close to the sources (especially iron, steel, and ferro-alloy industries). For Sb, on the other hand, it is likely that transport over long distances will occur. For each pair of metals, Table 6 compares the European emission ratios with concentration ratios for the late 1970s. Interesting features can be seen in the table. The Co/Mo ratio is similar for emissions and for the snow which is consistent with the rather similar size distributions for the two metals. The Co/Cr and Mo/Cr ratios are much larger for the snow than for the emissions, which is in good agreement with the above indication that Cr is found in significantly larger particles preventing their transportation far away from iron, steel, and ferro-alloy manufacturing industries which are located in low altitude valleys. Finally, the Co/Sb, Mo/ Sb, and especially Cr/Sb ratios are lower for the snow than for the emissions. This is consistent with the very small MMAD for Sb bearing particles which can then easily reach high altitude alpine areas far away from anthropogenic sources such as oil and coal fired power plants, fuel combustion, and refuse incineration. It is no doubt that the relative importance of the different source categories have strongly varied during the past centuries. This is clearly seen in Mont Blanc snow and ice when looking at the Cr/Sb and Mo/Sb concentration ratios, Figure 6. It is clear from the regression lines shown in Figure 6 that Cr/Sb and Mo/Sb ratios were significantly lower in snow and ice deposited before ∼1940s. This effect is especially clear for the Mo/Sb ratio. Co emissions nowadays mainly originate from fossil fuel combustion; there are indications that during previous centuries Co emissions were largely due to Cu and/or Ni smelter activities. Co is indeed often associated with Cu and Ni ores. Emission factors were probably large because of uncontrolled and wasteful smelting procedures (41, 42). Regarding Cr, the iron, steel, and ferro-alloy production has probably remained the main source of anthropogenic Cr during the last two centuries. The amounts produced were certainly smaller during the 18th and 19th centuries in Western Europe but increased rapidly during the following 3500
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 20, 1999
century (world production increased by a factor of ∼11 from 1900 to 1990 (43)). The iron and steel production was also quite important in the alpine valleys (44). For Mo, our data suggest that anthropogenic emissions were certainly very limited during the 18th and 19th centuries since at that time Mo found in Mont Blanc ice is largely natural. The increase in Mo concentrations observed in Mont Blanc ice after the ∼1940s is obviously linked with the ever increasing emissions from coal and oil burning. Finally for Sb if fossil fuel combustion is certainly an important contributor after the 1940s, it is likely that in the earlier periods there was a significant contribution from smelter activities. Sb is indeed found at rather high concentration levels in various ores which were smelted at that time using highly polluting technologies (45).
Conclusions This study has allowed for one of the first historical records of environmental contamination for Co, Cr, Mo, and Sb in midlatitude areas to be obtained. It shows that high altitude cold alpine snow and ice cores have a great potential to show regional trends in areas near human habitation. It will be interesting in the future to obtain new high altitude cold snow and ice cores taken from other locations in the Alps and other midlatitude mountain ranges. Especially, priority should be given to identifying drilling sites which have the potential of giving longer time series extending back to the Renaissance or Medieval times, without too much thinning of the annual layers along the depth profile. Also, great efforts should be given to various other metals of high interest such as Hg, V, and Ni.
Acknowledgments This work was supported in France by the Institut Universitaire de France, the Ministry of the Environment, the Agence de l’Environnement et de la Maıˆtrise de l’Energie, the Institut National des Sciences de l’Univers, and the University Joseph Fourier of Grenoble. In Italy, this study was performed in the framework of the Projects on Environmental Contamination and Glaciology and Paleoclimatology of the Italian Antarctic National Research Program. It was financially supported by ENEA through cooperation agreements with the Universities of Venice and Milan. Field sampling was supported by Re´gion Rhoˆne Alpes and the Ministry of the Environment (Environment Program of CNRS). We thank R. J. Delmas, S. Hong, A. Manouvrier, and C. Rado for their very kind participation in ice drilling, P. Cescon and G. Scarponi for their involvement in this study, and J. Jouzel and M. Stievenard for performing the deuterium measurements.
Literature Cited (1) Wolff, E. W.; Peel, D. A. Nature 1985, 313, 535. (2) Peel, D. A. In The Environmental Record in Glaciers and Ice Sheets; Oeschger, H., Langway, C. C., Eds.; John Wiley & Sons: New York, 1989; pp 207-223. (3) Raynaud, D.; Jouzel, J.; Barnola, J. M.; Chappellaz, J.; Delmas, R. J.; Lorius, C. Science 1993, 259, 926. (4) Wolff, E. W. Mar. Pollut. Bull. 1992, 25, 274. (5) Boutron, C. F. Environ. Rev. 1995, 3, 1. (6) Hong, S.; Candelone, J. P.; Patterson, C. C.; Boutron, C. F. Science 1994, 265, 1841. (7) Hong, S.; Candelone, J. P.; Patterson, C. C.; Boutron, C. F. Science 1996, 272, 246. (8) Candelone, J. P.; Hong, S. Pellone, C.; Boutron, C. F. J. Geophys. Res. 1995, 100, 16605. (9) Boutron, C. F.; Vandal, G. M.; Fitzgerald, W. F.; Ferrari, C. Geophys. Res. Lett. 1998, 25, 3315. (10) Murozumi, M.; Chow, T. J.; Patterson, C. C. Geochim. Cosmochim. Acta 1969, 33, 1247. (11) Wolff, E. W.; Suttie, E. D.; Peel, D. A. Atmos. Environ. 1999, 33, 1535. (12) Wolff, E. W.; Suttie, E. D. Geophys. Res. Lett. 1994, 21, 781.
(13) Barbante, C.; Turetta, C.; Capodaglio, G.; Scarponi, G. Int. J. Environ. Anal. Chem. 1997, 68, 457. (14) Hong, S.; Boutron, C. F.; Edwards, R.; Morgan, V. I. Environ. Res. 1998, 78, 94. (15) Van de Velde, K.; Boutron, C. F.; Ferrari, C.; Bellomi, T.; Barbante, C.; Rudnev, S.; Bolshov, M. Earth Planet. Sci. Lett. 1998,164, 521. (16) Rosman, K. J. R.; Ly, C.; Van de Velde, K.; Boutron, C. F. Earth Planet. Sci. Lett. In press. (17) Weiss, H. V.; Koide, M.; Goldberg, E. D. Science 1971, 172, 261. (18) Weiss, H. V.; Bertine, K.; Koide, M.; Goldberg, E. D. Geochim. Cosmochim. Acta 1975, 39, 1. (19) Boutron, C. F.; Leclerc, M.; Risler N. Atmos. Environ. 1984, 18, 1947. (20) Boutron, C. F. In Toxic Metals in the Atmosphere; Nriagu, J. O., Davidson, C.I., Eds.; John Wiley & Sons: New York, 1986; pp 467-505. (21) Schrauzer, G. N. In Metals and their Compounds in the Environment; Merian, E., Ed.; VCH: Weinheim, Germany, 1991; pp 879-892. (22) Gauglhofer, J.; Bianchi, V. In Metals and their Compounds in the Environment; Merian, E., Ed.; VCH: Weinheim, Germany, 1991; pp 852-878. (23) Davis, G. K. In Metals and their Compounds in the Environment; Merian, E., Ed.; VCH: Weinheim, Germany, 1991; pp 10891100. (24) Fowler, B. A.; Goering, P. L. In Metals and their Compounds in the Environment; Merian, E., Ed.; VCH: Weinheim, Germany, 1991; pp 742-750. (25) Nriagu, J. O.; Pacyna, J. M. Nature 1988, 333, 134. (26) Vincent, C.; Vallon, M.; Pinglot, F.; Funk, M.; Reynaud, L. J. Glaciol. 1997, 43, 513. (27) Candelone, J. P.; Hong, S.; Boutron, C. F. Anal. Chim. Acta 1994, 299, 9. (28) Boutron, C. F. Fresenius Z. Anal. Chem. 1990, 337, 482. (29) Paulsen, P. J.; Beary, E. S.; Bushee, D. S.; Moody, J. R. Anal. Chem. 1988, 60, 971.
(30) Barbante, C.; Bellomi, T.; Mezzadri, G.; Cescon, P.; Scarponi, G.; Morel, C.; Jay, S.; Van de Velde, K.; Ferrari, C.; Boutron, C. F. J. Anal. Atom. Spectrosc. 1997, 12, 925. (31) Barbante, C.; Cozzi, G.; Capodaglio, G.; Van de Velde, K.; Ferrari, C.; Boutron, C. F.; Cescon, P. J. Anal. Atom. Spectrosc. In press. (32) Shotyk, W.; Cheburkin, A. K.; Appleby, P. G.; Fankhauser, A.; Kramers, J. D. Earth Planet. Sci. Lett. 1996, 145, E1. (33) Nriagu, J. O Nature 1989, 338, 47. (34) Nriagu, J. O. In From Urban Air Pollution to Extra Solar Planets; Boutron, C. F., Ed.; EDP Sciences: Les Ulis, France, 1998; pp 205-218. (35) Wedepohl, K. H. Geochim. Cosmochim. Acta 1995, 59, 1217. (36) Duce, R. A. In Aerosol Forcing of Climate; Charlson, R. J., Heitzenberg J., Eds.; John Wiley & Sons: New York, 1995; pp 43-72. (37) Prange, A.; Kremling, K. Mar. Chem. 1985, 16, 304. (38) Arimoto, R.; Duce, R. A.; Ray, B. J.; Hewitt, A. D.; Williams, J. J. Geophys. Res. 1987, 92, 8465. (39) Pacyna, J. M. Atmos. Environ. 1984, 18, 41. (40) Quinn, T. L.; Ondov, J. M. Atmos. Environ. 1998, 32, 3467. (41) Hong, S.; Candelone, J. P.; Soutif, M.; Boutron, C. F. Sci. Total Environ. 1996, 188, 183. (42) Tylecote, R. F. In A History of Metallurgy; The Institute of Materials: 1992; pp 122-177. (43) Les Maıˆtres de l’Acier, Muse´e Dauphinois, Eds.; Grenoble, France, 1996. (44) Ten Decades of Mining Activity, Minesterio de Energia y Minas, Peru, information obtained in December 1998 from http:// www.mem.gob.pe. (45) Daumesnil, A. Maıˆtrise report; Geology Department, University Joseph Fourier: Grenoble, France, 1997; pp 12-21.
Received for review January 21, 1999. Revised manuscript received June 21, 1999. Accepted July 19, 1999. ES990066Y
VOL. 33, NO. 20, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3501